Abstract: A storage system is provided. The system includes a primary node having a processor and memory storing scheduling logic and a plurality of secondary nodes each of the secondary nodes having a processor and removable storage memory, wherein the processor of the primary node when executing the scheduling logic is configured to assign a priority to tasks executed by the storage system, monitor a processing load of the storage system, and monitor a capacity of the storage system. The processor is further configured to adjust power consumption of a processor of at least one of the plurality of secondary nodes based on the priority of tasks being executed by the storage system and the processing load and the capacity of the storage system.

Abstract: An amount of energy used by a storage device to complete shutdown procedures in response to a power loss by the storage device is determined. Performance parameters associated with a power loss protection (PLP) component of the storage device are received. A voltage provided to the PLP component to cause the PLP component to store an amount of energy that is greater than or equal to the amount of energy used by the storage device to complete the shutdown procedures is determined based on the performance parameters.

Abstract: A determination as to whether a workload of storage system operations to be performed by a storage system is below a threshold is made. In response to determining that the workload of the storage system operations is below the threshold, a first set of authorities is migrated from a first set of storage nodes of the storage system to a second set of storage nodes of the storage system. One or more processing devices previously executing the first set of authorities on the first set of storage nodes are in a reduced power mode.

Abstract: Buffering and persisting artificial intelligence (AI) training checkpoints, including: storing a plurality of artificial intelligence (AI) training checkpoints in a write buffer comprising a first storage memory type; determining that a condition for persisting an AI training checkpoint of the plurality of AI training checkpoints has been satisfied; and storing, in response to the condition being satisfied, the AI training checkpoint to persistent storage comprising a second storage memory type.

Abstract: A storage system is provided. The storage system includes a plurality of storage nodes configured to provide storage services for the storage system. A respective storage node comprises a processing device, a memory, network interfaces, and a plurality of device interfaces supporting hot-pluggable managed flash storage devices. A respective managed flash storage device comprises flash memory and a storage device controller supporting a set of commands for the respective storage node to access and optimize use of the flash memory for use in the storage system, wherein at least a subset of the managed flash storage devices further comprise memory and energy storage for staging writes, and for storing metadata for the storage system.

Abstract: Predictive device link health monitoring using machine learning, including: monitoring one or more lane margining metrics of a Peripheral Component Interface Express (PCIe) device; receiving, from a machine learning model, a failure prediction for the PCIe device based on an input to the machine learning model comprising the one or more lane margining metrics; and performing a remedial action associated with the PCIe device based on the failure prediction.

Abstract: Preserving data in a storage system operating in a reduced power mode, including: detecting that the storage system should enter the reduced power mode; and entering the reduced power mode, including performing, while in the reduced power mode, one or more maintenance operations on one or more storage devices in the storage system.

Abstract: Requests are received to write data into extents. The extents are logically contiguous address ranges having corresponding sizes that are larger than a minimum atomic write size of a storage system. A determination as to whether an amount of data written to a particular extent of the plurality of extents has reached a threshold is made. In response to determining that the amount of data written to the particular extent has reached the threshold, data written to the particular extent is stored in one or more managed flash storage devices utilizing a flash translation layer organized to store blocks of a size corresponding to the size of the extents.

Abstract: A determination as to whether a workload of storage system operations to be performed by a storage system is below a threshold is made. In response to determining that the workload of the storage system operations is below the threshold, a first set of authorities is migrated from a first set of storage nodes of the storage system to a second set of storage nodes of the storage system. One or more processing devices previously executing the first set of authorities on the first set of storage nodes are in a reduced power mode.

Abstract: A storage system is provided. The storage system includes a plurality of storage nodes configured to provide storage services for the storage system. A respective storage node comprises a processing device, a memory, network interfaces, and a plurality of device interfaces supporting hot-pluggable managed flash storage devices. A respective managed flash storage device comprises flash memory and a storage device controller supporting a set of commands for the respective storage node to access and optimize use of the flash memory for use in the storage system, wherein at least a subset of the managed flash storage devices further comprise memory and energy storage for staging writes, and for storing metadata for the storage system.

Abstract: A point-in-time copy of data of a source volume stored in a first set of blocks of a plurality of storage devices is generated. A command to relocate the point-in-time copy from the first set of data blocks of the plurality of storage devices to a second set of data blocks of the plurality of storage devices of a target volume while bypassing the storage system controller is transmitted by a storage system controller. The command includes address information associated with the second set of blocks. A mapping data structure is updated to associate read requests for the point-in-time copy with the second set of blocks of the target volume.

Abstract: A configuration of hardware resources of a storage system having multiple processing devices and multiple storage devices is received. A number of operations to be performed by the storage system over a range of time is predicted. Power usages of the storage system performing the plurality of operations using different performance settings of the hardware resources are estimated. Particular performance settings of the different performance settings to perform the operations are selected based on the estimated power usages.

Abstract: A determination as to whether a workload of storage system operations to be performed by a storage system is below a threshold is made. In response to determining that the workload of the storage system operations is below the threshold, a first set of authorities is migrated from a first set of storage nodes of the storage system to a second set of storage nodes of the storage system. One or more processing devices previously executing the first set of authorities on the first set of storage nodes are in a reduced power mode.

Abstract: Preserving data in a storage system operating in a reduced power mode, including: detecting that the storage system should enter the reduced power mode; and entering the reduced power mode, including performing, while in the reduced power mode, one or more maintenance operations on one or more storage devices in the storage system.

Abstract: A command to relocate data from a first set of data blocks at the plurality of storage devices to a second set of data blocks while bypassing the storage controller is transmitted by the storage controller. The command includes address information associated with the second set of blocks to store the relocated data.

Abstract: Writing data in a storage system that includes a first type of storage device and a second type of storage device, including: selecting, for one or more unprocessed write requests, a target storage device type from the first type of storage device and the second type of storage device; issuing a first group of write requests to the first type of storage device, the first group of write requests addressed to one or more locations selected in dependence upon an expected address translation to be performed by the first type of storage device; and issuing a second group of write requests to the second type of storage device, the second group of write requests addressed to one or more locations selected in dependence upon a layout of memory in the second type of storage device.

Abstract: Enabling communication between multiple storage controllers and a single-ported storage device, including determining, by an arbiter, that a first storage system controller of a plurality of storage system controllers has gained exclusive access to a single-ported storage device having a plurality of lanes; and in response to the determination, enabling communication between the first storage system controller and the storage device; and preventing communication between the storage device and at least one other storage system controller of the plurality of storage system controllers.

Abstract: A method to estimate remaining lifespan of a storage device is provided. The method includes obtaining first time series data for a set of metrics associated with operation of a set of flash storage devices, obtaining second time series data for one or more health metrics associated the set of storage devices, providing the first time series data for the set of metrics associated with operation of the plurality of flash storage devices and the second time series data for the one or more health metrics associated with the set of storage devices as training data to a machine learning model, and training the machine learning model to estimate a time to failure of a flash storage device based on the first time series data and the second time series data.

Abstract: A first command including data to be stored in non-volatile random access memory (NVRAM) of a storage device is received from a storage system controller by a storage device controller. In response to receiving the first command, the data is stored in the NVRAM. One or more second commands including a location in flash memory of the storage device for storing the data and an instruction to store the data in the location of the flash memory are received from the storage system controller. In response to receiving the one or more second commands, the data is stored in the location of the flash memory.

Abstract: Enabling communication between multiple storage controllers and a single-ported storage device, including determining, by an arbiter, that a first storage system controller of a plurality of storage system controllers has gained exclusive access to a single-ported storage device having a plurality of lanes; and in response to the determination, enabling communication between the first storage system controller and the storage device; and preventing communication between the storage device and at least one other storage system controller of the plurality of storage system controllers.